JPH04373085A - Background image replacing device - Google Patents

Abstract

PURPOSE:To obtain a background replacing image replacing background information with the other image at high speed by utilizing the information of difference between too still pictures, which are hourly continued, and the movement information of an object. CONSTITUTION:An inter-image subtraction executing part 21 in a differential image possessing part 2 obtains a differential image by executing a subtraction processing between a still picture (t-1) at time t-1 stored in an input image storage part 61 and a still picture (t) inputted at the time (t). Next, a movement information extraction part 42 obtains the movement information of objective information at the time (t) and according to this information, edge coordinate information is obtained. Finally, a background image replacing part 5 obtains the background replacing image replacing the background image out of the object at the time (t) with the image stored in a background image storage part 62 based on the edge coordinate information at the time (t).

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a background image substituting apparatus for obtaining a background substituting image by substituting a background image other than the object from a given still image with another previously prepared image.

0002

2. Description of the Related Art A method using difference information between two images has been proposed as a method for extracting a specific target area (for example, a person) from an image and replacing the background image. In this method, it is common to take and save a background image that does not include the target in advance, and then extract the target and replace the background image by taking the difference between this image and the image that does include the target. .

0003

[Problems to be Solved by the Invention] However, with the method of photographing the background in advance, there is a problem in that when the lighting conditions change or the camera position changes, the difference information cannot be used for target extraction.

[0004] An object of the present invention is to eliminate the need to take and save a background image in which no object is shown in advance.

[0005]

[Means for solving the problem] For still image signals that are input one after another, subtraction is performed between a still image (t-1) at time (t-1) and a still image t at time t. By taking the absolute value of the result, a difference image is obtained in which a moving area is extracted between the two still image signals. A binarized difference image is obtained by binarizing the difference image by threshold processing. Image processing such as contraction and expansion is performed on the binarized difference image to obtain a corrected binarized difference image in which noise is removed and edges are corrected. Y of the target at time t from the corrected binarized difference image
Vertex coordinate information regarding the coordinates and movement information indicating which direction the object has moved to the left or right are obtained. If it is determined from the movement information that the object has moved from left to right, the edge coordinates of the right half starting from the vertex coordinates of the object are the right edge coordinates starting from the vertex coordinates of the corrected binarized difference image. Using half edge coordinates, the edge coordinates of the left half starting from the vertex coordinates of the object are the edges of the left half of the object traced starting from the vertex coordinates of the object at time t in the still image at time t. Use coordinates. If it is determined from the movement information that the object has moved from right to left, the edge coordinates of the left half starting from the vertex coordinates of the object are the left half edge coordinates starting from the vertex coordinates of the corrected binarized difference image. Using half edge coordinates, the edge coordinates of the right half starting from the vertex coordinates of the target are the edges of the right half of the target traced starting from the vertex coordinates of the target at time t in the still image at time t. Use coordinates. The main feature is to obtain a background replacement image in which a background image other than the object at time t is replaced with another image based on the edge coordinate information of the object at time t obtained as described above.

[0006]

[Operation] By the above means, the present invention provides two temporally continuous
Using difference information between images and target movement information, it is possible to obtain a background replacement image in which a background image other than the target is replaced with another image prepared in advance. Then, in the still image at time t, a still image is obtained in which the background image other than the target is replaced with another image.

[0007]

[Example] The present invention will be explained in detail below using an example. FIG. 1 is a block diagram of one embodiment of the present invention. In the figure, 1 is an overall view of the background image replacement device, 2 is a difference image acquisition unit, 21 and 22 are inter-image subtraction execution units and difference image binarization units that constitute the difference image acquisition unit, and 3 is a motion Judgment unit, 4 is target area extraction unit, 41, 42, 43, 4
4 is an image correction section, a motion information extraction section, a differential image edge extraction section, and a target image edge extraction section, 5 is a background image replacement section, 6 is an image storage section, and 61, 62, and 63 are an input image storage section and a background, respectively. They are an image storage section and a background replacement image storage section, and are composed of these sections.

The operation of the embodiment will be explained below. When a still image t at time t is input to the background image replacement device 1, first, the inter-image subtraction execution unit 21 in the difference image acquisition unit 2
, subtraction is performed between the still image (t-1) at time (t-1) stored in the input image storage section 61 of the image storage section 6 and the still image t; A difference image is obtained by taking the absolute value of the subtraction result. This makes it possible to extract areas with movement between two still images.

Next, a difference image binarization unit 22 in the difference image acquisition unit 2 binarizes the difference image by threshold processing to obtain a binarized difference image. The threshold value is determined from the MAX gradation value to the adjacent gradation value on the density histogram, which is the sum of the frequencies of the gradation values every "5" for the binarized difference image to be subjected to threshold processing. A method is used in which the frequency difference with the value is checked and the gradation value immediately before the frequency difference becomes the maximum is set as the threshold value.

Next, the motion determining unit 3 checks the number of pixels with a gradation value of "1" in the binarized difference image, and if the number of pixels is less than a predetermined number, it is determined that there is no movement of the object between the two still images. It is determined that the amount is small, and the background replacement image obtained immediately before is used as the background replacement image, and the process moves to the next still image signal (t+1). If the number of pixels with a gradation value of "1" is greater than or equal to a predetermined number, it is determined that there is a large movement of the object between the two still images, and subsequent processing is performed.

When the binarized difference image is input to the target area extraction unit 4, first, the image correction unit 41 performs image processing such as contraction and expansion on the binarized difference image to eliminate noise in the binarized difference image. A corrected binarized difference image is obtained after removing and correcting edge breakage.

Next, the motion information extraction section 42 obtains vertex coordinate information of the object at time t and movement information indicating in which direction the object has moved, to the left or right. Extraction of the target vertex coordinates at time t is performed by extracting the vertex coordinates of the target from the still image at the time t at the time (
After subtracting the still image at t-1), obtaining an image consisting only of pixels for which the subtraction result is positive, and removing noise by contraction processing, move from the left corner (X coordinate = 0, Y coordinate = 0) to the right corner. It is checked whether there is a pixel with a gradation value of "1" toward , and the coordinates of the first detected pixel with a gradation value of "1" are taken as the coordinates of the top of the subject's head at time t. Regarding movement information, by comparing the vertex coordinates of the target at time t and the vertex coordinates of the target at time (t-1), it is determined whether the target has moved from left to right or from right to left. do.

If it is determined from the movement information that the object has moved from left to right, the difference image edge extraction unit 43 extracts the edge coordinates of the right half of the corrected binarized difference image starting from the vertex coordinates as the target. The edge coordinates of the right half of the target image are extracted as the edge coordinates of the right half starting from the vertex coordinates of the target image at time t, and the edges of the left half of the target are traced using the vertex coordinates of the target at time t as a starting point in the target image edge extraction unit 44. Extract the edge coordinates of the left half starting from the vertex coordinates. FIG. 2 shows an example of tracking the edge of the left half of a person from the coordinates of the top of the head after movement when the object is a person.

If it is determined from the movement information that the object has moved from right to left, the difference image edge extraction unit 43 extracts the edge coordinates of the left half of the corrected binarized difference image starting from the vertex coordinates as the object. The edge coordinates of the left half of the target image are extracted as the edge coordinates of the left half starting from the vertex coordinates of the target image, and the edge coordinates of the right half of the target are traced in the target image edge extraction unit 44 with the vertex coordinates of the target at time t as the starting point. Extract the edge coordinates of the right half starting from the vertex coordinates.

Next, the background image replacement unit 5 stores a background image other than the target at time t in the background image storage unit 62 in the image storage unit 6 based on the edge coordinate information of the target at time t obtained as described above. Replace the image with the original image to obtain a background replacement image. This background replaced image is output from the background image replacing device 1 and is stored in the background replaced image storage section 63 in the image storage section 6.

[0016]

[Effects of the Invention] As explained above, according to the present invention,
Since it uses the difference information between two temporally consecutive still images and the movement information of the target, there is no need to take and save a background image that does not include the target in advance. It becomes possible to quickly obtain a background replacement image in which a background image other than the target image is replaced with another image without being affected. Therefore, if the present invention is applied,
For example, when communicating using a still image TV phone, it is possible to implement an additional service that sends a different image as the background in response to a user's request to not show the background of anything other than the person on the TV. This has the effect of promoting the spread of the use of image communication services by telephone and the like.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing one embodiment of the present invention.

FIG. 2 is a diagram illustrating a method of tracing edge coordinates of an object.

[Explanation of symbols]

1 Background image replacement device 2 Difference image acquisition unit 21 Inter-image subtraction execution unit 22 Difference image binarization unit 3 Motion determination section 4 Target area extraction part 41 Image correction section 42 Motion information extraction section 43 Difference image edge extraction part 44 Target image edge extraction section 5 Background image replacement part 6 Image storage section 61 Input image storage section 62 Background image storage section 63 Background replacement image storage section

Claims (1)

[Claims] Claim 1: Perform subtraction between a still image (t-1) at time (t-1) and a still image t at time t for still images that are input one after another, and take the absolute value of the subtraction result. means for obtaining a difference image by extracting a moving area between the two still images; means for obtaining a binarized difference image by binarizing the difference image by threshold processing; and a corrected binarized difference image. means for obtaining vertex coordinate information regarding the Y-coordinate of the object at time t and movement information indicating which direction the object has moved from left to right, and when it is determined from the movement information that the object has moved from left to right; As the edge coordinates of the right half starting from the vertex coordinates of the target, use the edge coordinates of the right half starting from the vertex coordinates of the corrected binarized difference image,
As the edge coordinates of the left half starting from the vertex coordinates of the object, use the edge coordinates of the left half of the object traced from the vertex coordinates of the object at time t as the starting point in the still image at time t, and If it is determined that the object has moved from right to left, the edge coordinates of the left half starting from the vertex coordinates of the object are the edges of the left half starting from the vertex coordinates of the corrected binarized difference image. Using coordinates,
means for using the edge coordinates of the right half of the object traced from the vertex coordinates of the object at time t as a starting point in a still image at time t as the edge coordinates of the right half starting from the vertex coordinates of the object; A background characterized by comprising means for obtaining a background replacement image in which a background image other than the object at time t is replaced with another image prepared in advance based on the edge coordinate information of the object at time t obtained by. Image replacement device.